Much of our recent growth regulation research has been concerned with the ras oncogene and its positive regulators, the Ras-specific guanine nucleotide exchange factors (Ras-GNEFs). These molecules play a key role in signal transduction, and ras is mutationally activated in many human and animal tumors. The analysis of two closely related, widely expressed Ras-GNEFs, sos1 and sos2, has identified important differences in the stability of their encoded proteins and in the duration of their signaling properties. The Sos2 protein carries out long-term signaling less efficiently than Sos1 to a degree that physiologic levels of Sos2 cannot support transformation by oncoproteins such as tyrosine kinases, which rely on Sos-dependent activation of Ras for their transforming activity. In addition, the Sos2 protein is a physiologic substrate for degradation by the ubiquitin-proteasome system, while Sos1 is not subject to such regulation. We are currently examining the genetic basis for the difference in Sos1 vx. Sos2 signaling, as well as the regulation of Sos1 function. We have found that E-cadherin can interfere with the signaling function of Sos, probably via inhibition of upstream molecules. The implications for inhibition of transformation by E-cadherin are under investigation. We are also studying GRF1 and GRF2, two closely related Ras-GNEFs with more restricted expression than Sos which are activated by calcium. We have found that the calcium-dependent activation of GRF stimulates Raf via two signals, one that is Ras-dependent and another that is Ras-independent. Also, GRF1 and GRF2 form homo- and hetero-dimers through their Dbl domains. Dimerization may be required for propagating the Ras-independent signal from GRF that stimulates Raf. Unexpectedly, we also find that Ras-GRF stimulates the Ras-related protein Rap1a, probably via the Ras catalytic region of GRF.